| LeePen | onefang: Just a heads up that I am removing some cruft from dak which will propagate /devuan on the mirrors. | 13:14 |
|---|---|---|
| onefang | So that means less disk space needed by mirrors? | 13:14 |
| LeePen | Probably. Not sure how much though. | 13:16 |
| golinux | I have started a pad to collect a list of new behaviors for inclusion in the Chimaera Release Notes: https://pad.dyne.org/code/#/2/code/edit/zrwXb4YnOlmO6fy2mPq4AsmR/ | 21:49 |
| fsmithred_ | mason: https://dev1galaxy.org/viewtopic.php?pid=22455#p22455 | 22:53 |
| mason | fsmithred_: Right, but that's talking about a higher level. | 22:54 |
| mason | fsmithred_: The root issue is that the cryptsetup teardown can't happen. | 22:55 |
| mason | Anyway, replied asking what you think about the crypttab idea. | 22:55 |
| fsmithred_ | I guess I don't understand the problem. What is the result of no teardown? | 22:55 |
| mason | fsmithred_: No real impact. | 22:56 |
| mason | fsmithred_: It's just notionally unclean. BUt it's what all the patches amount to. | 22:56 |
| mason | The issue we're patching is the recursive back-off as the system tries the teardown regardless of the fact that it can't succeed. | 22:57 |
| fsmithred_ | yeah, that subject comes up from time to time. | 22:57 |
| fsmithred_ | I think the hypothetical solution is to switch to initramfs on shutdown. | 22:57 |
| fsmithred_ | at least that's a suggestion that's been floated | 22:58 |
| mason | Right, but shutdown-ramfs is more notional than otherwise. It *is* the right solution. | 22:58 |
| fsmithred_ | I think that change is above my pay grade. | 22:58 |
| mason | I can implement it if you like it. What I want is an opinion. :) | 22:59 |
| fsmithred_ | my opinion is to leave stuff alone if it works | 22:59 |
| mason | Gotcha. It doesn't work which is why we're patching it, but that's enough directive for me to run with. | 22:59 |
| fsmithred_ | only reason I got involved with the cryptsetup patch was to deal with the shutdown delay | 22:59 |
| fsmithred_ | and that seems to be fixed upstream | 23:00 |
| mason | for sysvinit? how so? | 23:03 |
| fsmithred_ | there is no longer a shutdown delay with encrypted lvm | 23:03 |
| fsmithred_ | I'll test encrypted root without lvm later | 23:04 |
| mason | Ah, you're suggesting that what I see is probably user error? | 23:04 |
| mason | I mean, I see it every shutdown, and it matches the expected behaviour, but I guess I could be confused. | 23:04 |
| fsmithred_ | you're seeing the delay in beowulf? | 23:04 |
| mason | No, Chimaera. Beowulf you noted that we're shipping the patch in that modified package. | 23:04 |
| mason | Without that package, I do see the breakage, which is why I was using my patch and a divert. | 23:05 |
| fsmithred_ | well, you have to install the patch file manually. | 23:05 |
| mason | How's that fixed upstream then? I don't understand. | 23:05 |
| fsmithred_ | in chimaera | 23:05 |
| mason | It's definitely not fixed in Chimaera, as noted. | 23:05 |
| fsmithred_ | if you're seeing the problem in chimaera, I'd like to be able to reproduce it. | 23:05 |
| mason | What's it got to do with LVM, anyway? | 23:06 |
| mason | The problem is in cryptsetup. | 23:06 |
| fsmithred_ | there were two problems | 23:06 |
| mason | Ah. I don't run LVM, so that one doesn't apply. | 23:06 |
| fsmithred_ | delay with encrypted root and delay with lvm | 23:06 |
| fsmithred_ | devujan provided the part that fixed lvm | 23:06 |
| fsmithred_ | I don't recall who first suggested getting rid of the timeout loop | 23:07 |
| fsmithred_ | (but I'm glad they did) | 23:07 |
| mason | Dunno. That's what I've been doing anyway. It's the solution, so probably several of us hit it at once. | 23:07 |
| fsmithred_ | as I said, I'll do an install without the lvm and see what happens. | 23:07 |
| fsmithred_ | did you say you're using zfs with this? | 23:08 |
| mason | I've seen it elsewhere. All you need is LUKS under root. | 23:09 |
| fsmithred_ | well, that's the next test | 23:09 |
| fsmithred_ | and I just realized I used the wrong iso to test the grub-theme fix. :( | 23:14 |
| golinux | Oh-oh | 23:19 |
| golinux | I don;t know how you keep all the moving parts straight . . . | 23:20 |
| fsmithred_ | look at one thing at a time | 23:23 |
| mason | yar | 23:24 |
| Centurion_Dan | Hi! | 23:26 |
| golinux | Hello stranger | 23:27 |
| fsmithred_ | hi | 23:27 |
| Centurion_Dan | I got given a laptop to fix today.... one I haven't seen in a very long time. | 23:29 |
| Centurion_Dan | Apparently emails stopped working recently.... on the Linux side (it's dual boot Win10 and Linux) | 23:29 |
| Centurion_Dan | It's running Devuan Jessie. | 23:30 |
| Centurion_Dan | Will it upgrade to Chimaera?? | 23:31 |
| fsmithred_ | maybe | 23:31 |
| fsmithred_ | I think ascii to chimaera has been done | 23:31 |
| fsmithred_ | or else it was jessie to beowulf | 23:31 |
| Centurion_Dan | I'll step up to asciil and maybe do ascii to chimaera.. | 23:49 |
| golinux | That might be the safest way | 23:50 |
| fsmithred_ | no | 23:52 |
| fsmithred_ | probablly will cause problems | 23:52 |
| fsmithred_ | what's the version of eudev in ascii? | 23:52 |
| golinux | Oh I see that beowulf wan't in the line | 23:54 |
| fsmithred_ | mason, I just did encrypted root install. Shutdown took three seconds. | 23:54 |
| fsmithred_ | golinux are you running ascii? | 23:54 |
| mason | fsmithred_: I guess you'd accept a screenshot or something? I can take a photo so you'll believe it's still an issue? | 23:54 |
| fsmithred_ | I don't disbelieve you | 23:55 |
| fsmithred_ | I want to know what the difference is | 23:55 |
| golinux | fsmithred_: You know I'm still on jessie that was installed by Centurion_Dan all those years ago. | 23:56 |
| golinux | I am running ascii on the other machine but not touched it for almost 2 weeks | 23:56 |
| golinux | I really don't like to have to rebuild my work environment | 23:57 |
| fsmithred_ | ok, eudev is 3.2 in ascii. You're save. | 23:58 |
| fsmithred_ | safe | 23:58 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!